Article: "Even Our Willingness to Fly is Partisan"

Thank you, Dave. As you noted, the statistical analysis is quite interesting, and the regional differences telling.

Interesting article. Last week I flew CLT-SFO and back. SFO was empty but then its a hub for international travelers while Charlotte was packed but it's an east coast hub. Unrelated to the article, alcohol was not served in coach (this was a 5 hour non-stop). I wonder if this would cut down "anger issues" on flights.

Interesting but not surprising. Alcohol, in my opinion, would only exacerbate the "anger".

I've read it three times, and still don't understand what it says that is significant. But mainly I object to speaking of being in a "recovery" as if the underlying cause of the economic problems (the pandemic) is over and we can expect no further issues impacting travel.

Dave, this topic will probably get enough comments that drift into the political sides to get pulled, but in the meantime, watch any conclusions made on regression lines with such low R-squares numbers.

@stan... Recovery, in this sense, just means the recovery of passengers flying again. It does not mean there are not going to be ups and downs (including potentially significant downs)... just like a patient recovering from a heart attack can have ups and downs (including death!).

@Jean... Agreed, the coefficient of determination shows a mild effect... definitely one that would not justify the "hyper-partisan" phrase that appears in the URL, but not in the headline of the article on the website. Then there is also the matter of spurious correlations... for real fun, look at this spurious correlations website... https://tylervigen.com/old-version.html

The correlation between the annual "Number people who drowned by falling into a swimming-pool" and "Number of films Nicolas Cage appeared in" has a much stronger correlation coefficient than the one in this article.

Re: the low r-squared...

It's Sunday, and I'm not going to dig out my stats text, so going off memory here and could be completely wrong...

I counted 147 data points in the first graph, which means the degrees of freedom is going to be large. This will make it much more likely that low r-squared figure is statistically significant.

And, sorry if I'm overthinking this Dave :-)

Eric,

I'd rather you overthink the statistics than the politics. 🤓

I don't feel that the text makes it clear enough whether the charts reflect appropriate adjustments for issues like lower population densities in redder areas. The author acknowledges this issue, but her discussion of the statistics lacks academic rigor.

I bet Georgia's vote in the recent election doesn't decrease traffic at Hartsfield, especially since those of us in nearby cities no longer have flights anywhere else.

This reminds me of the meme that was circulating last month that went something like

"I'm not sayin' nothin' but everyone who confuses correlation with causation eventually ends up dead."
Ha.

everyone who confuses correlation with causation eventually ends up dead.

I think everyone who doesn't confuse correlation with causation also eventually ends up dead, don't they.

the annual "Number people who drowned by falling into a swimming-pool" and "Number of films
Nicolas Cage appeared in" has a much stronger correlation coefficient

I doubt you can prove that.

I remember in the 60s, when people who were told that lung cancer had a strong correlation with smoking countered that the increase in lung cancer correlated with the number of bath tubs in the United States. Most of those people died in the next few years; I knew some of them.

There are three kinds of lies: lies, damned lies, and statistics.

98% of people who say that have just been shown statistics that contradict beliefs they hold without any supporting evidence.

Another interesting statistics is that, since the population of the earth has been doubling every 30 or so years for millennia, half of the people who have ever lived on the earth are alive today. So, that makes the death rate 50%, doesn't it? Does that mean that everyone has a 50% chance of living forever?

Here is how you will know when the airlines are recovering......your email inbox will stop being inundated with amazing flight deals, reminders of less change rules and regulations, and the "we care about you" messages.

correlation vs causation

True, a casual statistical relationship (like cancer and bath tubs) does not prove causation. However, when a cause/effect mechanism has been postulated (like political orientation makes people more likely to ignore the risk of the virus and fly), and it is supported by a statistical relationship, then it is certainly strong evidence for the postulated causation.

I postulated that because Republicans tend to forego masks and distancing because they believe their leader when he claims those things aren't effective, they are spreading the disease, and their states will have more new cases.

Every day, the CDC publishes a counting of the new cases and deaths for the previous 7 days. I've been tracking these numbers for months; the last date for which I recorded them was Nov 12.

On that date, the average number of new cases for the whole country was 265 per 100k population. I plotted the % vote by state for the Republican candidate in 2016 (the 2020 vote count is not yet validated) vs the number of new cases and saw a trend towards an increase in the number of positive cases as the Republican percent went up. The average number was 300 for red states and 220 for blue states. Fifteen out of 21 blue states (71%) were below the average, while only 11 of the 30 red states (37%) were below average.

I did a regression analysis for the date. Turns out the correlation between Republican vote % and new cases was a 50% positive correlation. I think this analysis tends to support the findings (above) about flying vs political orientation.

teeters? How about dives in? Unless you want to make it about the science of statistics, then the subject is irrelevant and has no relevance to travel. But all that aside. Its interesting in a political sort of way. Thanks for posting. Enjoyed it.

In the before-times, was willingness to fly related to prices and convenient flight times and airport connections? Was (is) there a correlation between political affiliation and income level (maybe hence, price sensitivity) and schedule availability and convenient access to major airports?

I think there has always been an association between price sensitivity (related to income level) and convenience, but I think this thing with air travel recovery vs politics is a recent phenomenon, and related to the risk taking inherent to political orientation evident in the increase in infection seen in some states.

And I do think this discussion is relevant to travel in that, if we ever want the Europeans to let us go there again, we have to get this virus under control in our country. Not to mention that fact that they are having their own problems, and we might not even want to go there soon.

BTW, I can't agree with the linked article that shows Denver as a huge red dot of "Originating passenger recovery". The last time I was at DIA (less than two months ago), it was a ghost town, at least in ticketing and baggage claim. Now, it might still be an active hub for through passengers, but I didn't see many originating or terminating passengers. And we are pretty far blue in the political spectrum.